There is a question quietly unsettling everyone who makes a living in front of or behind an Australian camera, and it is not about box office or streaming budgets. It is about ownership. When a machine can reproduce an actor’s face, clone a voice from a few seconds of dialogue, or generate a script in the style of a working writer, the industry is forced to ask something that once seemed absurd: who actually owns you?
That is the provocation at the heart of a recent analysis published by the screen trade outlet FilmInk, which examines how generative artificial intelligence is colliding with the rights, livelihoods and creative identities of people who work in Australian film and television. You can read the original piece, Who Owns You? Australian Screen in the Age of AI, at FilmInk. The framing is deliberately blunt, because the stakes for a comparatively small industry are anything but abstract.
The context: a sector already under pressure
Australia’s screen industry has spent the past decade fighting for oxygen. It competes for talent and finance against Hollywood, contends with the buying power of global streamers, and leans heavily on a mix of government incentives, local content obligations and the sheer stubbornness of its practitioners. Into that fragile ecosystem has arrived a technology that can, at least in narrow tasks, imitate the very things the industry sells: a performance, a likeness, a turn of phrase, a distinctive visual style.
The anxiety is not hypothetical. Tools that generate video from text prompts have improved dramatically, voice cloning is cheap and convincing, and digital doubles of performers are already used in productions around the world. What began as a special-effects convenience has become a question of principle. If a studio can license an actor’s scanned likeness once and then deploy it indefinitely, what happens to the working actor who used to be paid per job? If a model has been trained on decades of Australian scripts and films, who benefits when it spits out something that feels familiar?
The news: rights, consent and the training-data problem
The through-line of the FilmInk piece is that Australian law was not built for this moment. Copyright protects specific works, but it sits awkwardly with the idea of protecting a person’s likeness or voice as such, and it offers little comfort when the grievance is that a model was trained on your work without permission. Performers have some contractual protections, but contracts are only as strong as the bargaining power behind them, and individual freelancers rarely hold the leverage to dictate terms to a large producer or a technology platform.
This is why the conversation keeps circling back to consent, compensation and control, the three principles that unions and guilds have pushed to the centre of the debate. The Media, Entertainment and Arts Alliance, the union covering Australian performers and journalists, has been vocal about the need for informed consent before a performer’s image or voice is digitally replicated, ongoing payment when that digital version is used, and meaningful say over how it is deployed. Those demands echo the terms that United States performers won after the prolonged 2023 strikes, and they are now shaping enterprise bargaining and production agreements here.
Two views: existential threat, or overdue reckoning
There are, broadly, two ways the industry is reading this moment. The first is defensive and fearful. On this view, generative AI is an extractive technology that has already helped itself to a vast body of human creative work without licence or payment, and is now positioned to undercut the people it learned from. Actors worry about being scanned into obsolescence, writers worry about being reduced to editors of machine drafts, and crew worry that budget savings from automation will flow upward rather than to the workforce.
The second view is more pragmatic and, in places, cautiously optimistic. AI can lower the cost of visual effects, previsualisation and post-production, which in a chronically underfunded sector could mean more projects actually getting made. Emerging filmmakers gain access to tools that were once the preserve of big-budget productions. On this reading the technology is not the enemy so much as an untamed force that needs rules, and the task is to build a licensing and consent framework that pays creators fairly while letting the industry use the tools that everyone else in the world will be using anyway. Both camps tend to agree on one thing: doing nothing is the worst option.
The Australian stakes
For Australia specifically, the ownership question lands harder because scale offers no cushion. A large market can absorb disruption and still sustain a domestic industry through sheer volume. A smaller one cannot, which means the loss of even a slice of paid work, or the erosion of the incentive to develop original local stories, could hollow out capability that took generations to build. There is also a cultural dimension. Australian screen work is part of how the country tells itself its own stories, and a future in which those stories are increasingly assembled by models trained on someone else’s back catalogue raises questions that go well beyond economics.
This is unfolding against a broader national reckoning with AI policy. The federal government has been wrestling with how, and whether, to regulate AI, including contentious questions about copyright and whether tech companies should be free to train models on Australian creative work. Creative advocates have pushed for a seat at the table as the country stands up new AI institutions, arguing that artists and performers cannot be an afterthought in decisions that will reshape their industries. The screen sector’s fight over ownership is one front in that much larger campaign.
What’s next
The immediate battleground is contracts and collective agreements, where consent and compensation clauses covering digital replicas are being negotiated production by production. Beyond that lies the slower, harder work of law reform, where Australia will have to decide whether existing copyright and moral-rights frameworks can be stretched to cover likeness and training data, or whether new protections are needed. Government reviews of AI and copyright will be watched closely by everyone in the sector, because the answers will determine whether performers and writers keep a grip on the things that make them individually valuable.
For now the question in the FilmInk headline stands unanswered, which is precisely the point. Ownership of your face, your voice and your creative output used to be a given. In the age of AI it has become something Australian screen workers may have to fight to keep, one clause and one policy at a time.
Sources: FilmInk.
